The Computation Problem with Sequential Decoding
نویسنده
چکیده
Fano Sequential Decoding is a technique for communicating at a high information rate and with a high reliability over a large class of channels. However, equipment cost and variation in the time required to decode successive transmitted digits limit its use. This work is concerned with the latter limitation. Others have shown that the average processing time per decoded digit is small if the information rate of the source is less than a rate Rcomp. This report studies the probability distribution of the processing time random variable and applies the results to the buffer overflow probability, i.e., the probability that the decoding delay forces incoming data to fill and overflow a finite buffer. It is shown that the overflow probability is relatively insensitive to the buffer storage capacity and to the computational speed of the decoder, but quite sensitive to information rate. In particular, halving the source rate more than squares the overflow probability. These sensitivities are found to be basic Sequential Decoding and arise because the computation per decoded digit is large during an interval of high channel noise and grows exponentially with the length of such an interval. A conjecture is presented concerning the exact behavior of the overflow probability with information rate. This conjecture agrees well with the (limited) experimental evidence available.
منابع مشابه
Sequential decoding for lossless streaming source coding with side information
The problem of lossless fixed-rate streaming coding of discrete memoryless sources with side information at the decoder is studied. A random time-varying tree-code is used to sequentially bin strings and a Stack Algorithm with a variable bias uses the side information to give a delay-universal coding system for lossless source coding with side information. The scheme is shown to give exponentia...
متن کاملLower and Upper Bounds on the Distribution of Computation for Sequential Decoding of Specific Linear Tree Codes.
متن کامل
Coding and decoding for time-discrete amplitude-continuous memoryless channels
In this report we consider some aspects of the general problem of encoding and decoding for time-discrete, amplitude-continuous memoryless channels. The results can be summarized under three main headings. 1. Signal Space Structure: A scheme for constructing a discrete signal space, for which sequential encoding-decoding methods are possible for the general continuous memoryless channel, is des...
متن کاملHardware Architecture for Modified Sequential LDPC Decoder
Iterative decoding of Low Density Parity Check (LDPC) codes using the Parity Likelihood Ratio (PLR) algorithm have been proved to be more efficient compared to conventional Sum Product Algorithm (SPA). However, the nature of PLR algorithm tends to put numerious pieces of data to this decoder and perform computation intensive operations, which is a major challenge for building a practical real-t...
متن کامل